RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-26 19:27:45 +00:00
Code Issues Activity Actions

LibJS: Resolve all UTF-8 rope strings as a String

Browse source
This commit is contained in:
Timothy Flynn 2023-01-14 10:31:18 -05:00 • committed by Linus Groh
parent 58bc831750
commit 46dd8c1c0b
1 changed files with 11 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

19
Userland/Libraries/LibJS/Runtime/PrimitiveString.cpp
View file

@ -275,21 +275,24 @@ ThrowCompletionOr<void> PrimitiveString::resolve_rope_if_needed() const
for (auto const* current : pieces) { for (auto const* current : pieces) {
if (!previous) { if (!previous) {
// This is the very first piece, just append it and continue. // This is the very first piece, just append it and continue.
TRY(builder.append(TRY(current->deprecated_string()))); TRY(builder.append(TRY(current->utf8_string())));
previous = current; previous = current;
continue; continue;
} }
// Get the UTF-8 representations for both strings. // Get the UTF-8 representations for both strings.
auto previous_string_as_utf8 = TRY(previous->deprecated_string()); auto current_string_as_utf8_string = TRY(current->utf8_string());
auto current_string_as_utf8 = TRY(current->deprecated_string()); auto current_string_as_utf8 = current_string_as_utf8_string.bytes_as_string_view();
auto previous_string_as_utf8_string = TRY(previous->utf8_string());
auto previous_string_as_utf8 = previous_string_as_utf8_string.bytes_as_string_view();
// NOTE: Now we need to look at the end of the previous string and the start // NOTE: Now we need to look at the end of the previous string and the start
// of the current string, to see if they should be combined into a surrogate. // of the current string, to see if they should be combined into a surrogate.
// Surrogates encoded as UTF-8 are 3 bytes. // Surrogates encoded as UTF-8 are 3 bytes.
if ((previous_string_as_utf8.length() < 3) || (current_string_as_utf8.length() < 3)) { if ((previous_string_as_utf8.length() < 3) || (current_string_as_utf8.length() < 3)) {
TRY(builder.append(TRY(current->deprecated_string()))); TRY(builder.append(current_string_as_utf8));
previous = current; previous = current;
continue; continue;
} }
@ -297,7 +300,7 @@ ThrowCompletionOr<void> PrimitiveString::resolve_rope_if_needed() const
// Might the previous string end with a UTF-8 encoded surrogate? // Might the previous string end with a UTF-8 encoded surrogate?
if ((static_cast<u8>(previous_string_as_utf8[previous_string_as_utf8.length() - 3]) & 0xf0) != 0xe0) { if ((static_cast<u8>(previous_string_as_utf8[previous_string_as_utf8.length() - 3]) & 0xf0) != 0xe0) {
// If not, just append the current string and continue. // If not, just append the current string and continue.
TRY(builder.append(TRY(current->deprecated_string()))); TRY(builder.append(current_string_as_utf8));
previous = current; previous = current;
continue; continue;
} }
@ -305,7 +308,7 @@ ThrowCompletionOr<void> PrimitiveString::resolve_rope_if_needed() const
// Might the current string begin with a UTF-8 encoded surrogate? // Might the current string begin with a UTF-8 encoded surrogate?
if ((static_cast<u8>(current_string_as_utf8[0]) & 0xf0) != 0xe0) { if ((static_cast<u8>(current_string_as_utf8[0]) & 0xf0) != 0xe0) {
// If not, just append the current string and continue. // If not, just append the current string and continue.
TRY(builder.append(TRY(current->deprecated_string()))); TRY(builder.append(current_string_as_utf8));
previous = current; previous = current;
continue; continue;
} }
@ -314,7 +317,7 @@ ThrowCompletionOr<void> PrimitiveString::resolve_rope_if_needed() const
auto low_surrogate = *Utf8View(current_string_as_utf8).begin(); auto low_surrogate = *Utf8View(current_string_as_utf8).begin();
if (!Utf16View::is_high_surrogate(high_surrogate) || !Utf16View::is_low_surrogate(low_surrogate)) { if (!Utf16View::is_high_surrogate(high_surrogate) || !Utf16View::is_low_surrogate(low_surrogate)) {
TRY(builder.append(TRY(current->deprecated_string()))); TRY(builder.append(current_string_as_utf8));
previous = current; previous = current;
continue; continue;
} }
@ -328,7 +331,7 @@ ThrowCompletionOr<void> PrimitiveString::resolve_rope_if_needed() const
previous = current; previous = current;
} }
m_deprecated_string = builder.to_deprecated_string(); m_utf8_string = TRY(builder.to_string());
m_is_rope = false; m_is_rope = false;
m_lhs = nullptr; m_lhs = nullptr;
m_rhs = nullptr; m_rhs = nullptr;